- Email: [email protected]
Formation of hydroxyapatite on CaSiO3 powders in simulated body fluid
Journal of the European Ceramic Society 22 (2002) 511–520 www.elsevier.com/locate/jeurceramsoc
Formation of hydroxyapatite on CaSiO3 powders in simulated body fluid Punnama Siriphannona,1, Yoshikazu Kameshimaa, Atsuo Yasumoria, Kiyoshi Okadaa,*, Shigeo Hayashib a Department of Inorganic Materials, Tokyo Institute of Technology, O-okayama, Meguro, Tokyo 152-8552, Japan Research Institute of Materials and Resources, Mining College, Akita University, Tegatagakuen, Akita 010-8502, Japan
b
Received 10 January 2001; received in revised form 10 April 2001; accepted 21 April 2001
Abstract CaSiO3 powders were prepared from ethanol solutions of Ca(NO3)2.4H2O and Si(OC2H5)4 using NaOH as a precipitant. The resultant powders were heated at three different temperature regimes, (1) 500 C, (2) 500 and 1000oC and (3) 500 and 1400 C, to obtain the amorphous phase (amorphous-CS), low temperature phase (b-CS), and high temperature phase (a-CS) of CaSiO3, respectively. The different amorphous and crystalline phases exhibited different microtextures and specific surface areas of the powders. The rough, porous particles of amorphous-CS and b-CS have higher specific surface areas than the smooth, dense particles of a-CS. These CaSiO3 powders were soaked in a simulated body fluid (SBF) at 36.5 C for 2 h to 30 days. Formation of hydroxyapatite (HAp) was observed on the surfaces of all samples, but the formation behavior and microstructures were different, resulting the differences in microstructure and crystal structure of the starting powders as well as particle size and specific surface area. The HAp formed on the amorphous-CS was a loose porous layer consisting of uniformly-sized tiny ball-like agglomerated particles, while that formed on the b-CS and a-CS was a dense layer consisting of larger ball-like agglomerated particles. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Apatitie; Bioactive materials; CaSiO3 powders; Simulated body fluid (SBF)
1. Introduction During the last four decades, various materials known as ‘bioactive materials’ such as glasses,1 3 sintered hydroxyapatite,4 glass ceramics,5 7 composite materials,8 10 etc., have been synthesized and developed for medical applications. A significant characteristic of bioactive materials is their ability to bond with living bone through the formation of a hydroxyapatite (HAp) interface layer. The same HAp layer is reported to form on the surfaces of bioactive materials after soaking in a biomimetic system of simulated body fluid (SBF) which has ion concentrations similar to human blood plasma. The proposed mechanism of HAp formation involves the dissolution of calcium ions from the bioactive materials into the body fluid, with the simultaneous formation of * Corresponding author. E-mail address: [email protected] (K. Okada). 1 Present address: Department of Chemistry, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Chalongkrung Road, Ladkrabang, Bangkok 10520, Thailand.
a silica-rich interlayer on the material surfaces. The dissolution of the calcium ion increases the degree of supersaturation of the surrounding fluid with respect to HAp, while the silica-rich interlayer provides favorable sites for HAp nucleation by dissolving an appreciable amount of silicate ion.11 13 The nucleation and growth of the HAp layer proceeds by reaction of the calcium, phosphate, and hydroxide ions from the surrounding fluid,2,3,13 15 sometimes incorporating with carbonate or fluoride anions.2,14,16 19 a-CaSiO3 ceramics prepared from coprecipitated powders using NaOH as the precipitant have been proposed as new candidate bioactive materials by the present authors since they can induce the fast formation of an HAp layer after soaking in SBF solution.20 22 It is believed that the presence of calcium and silicate ions in the a-CaSiO3 ceramics plays an important role in the formation of HAp.23 The present study investigates the effect of the structure of the CaSiO3 powders on the formation behavior of HAp by soaking various amorphous and crystalline CaSiO3 powders in a SBF solution for varying periods.
0955-2219/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0955-2219(01)00301-6
512
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
and dried at room temperature. The soaking experiment was done only once for amorphous-CS, b-CS and a-CS, respectively.
2. Experimental 2.1. Preparation of CaSiO3 powders CaSiO3 powder was prepared by coprecipitation, using Ca(NO3)2.4H2O and Si(OC2H5)4 (TEOS) as the raw materials. These reagents were dissolved in 500 ml of ethanol, the concentration was adjusted to 0.2 mol/l, and the solution was stirred for 2 h. Precipitation was obtained by adding 0.33 mol/l NaOH solution to the solution. The precipitated gel was filtered, washed once with distilled water, dried in an oven at 100 C overnight, and calcined at 500 C for 2 h. The chemical composition of the as-prepared powder determined by X-ray fluorescence (XRF; RIX3000, Rigaku Co., Tokyo, Japan) was CaO 46.2, SiO2 53.4, and Na2O 0.4 mass% on a dryweight basis. The calcined powder (amorphous CaSiO3) was divided into three parts and subjected to different firing processes as follows; (1) fired at 1000oC for 2 h to obtain the low temperature crystalline phase b-CaSiO3 and (2) fired at 1400oC for 2 h to obtain the high temperature phase aCaSiO3. The resultant powders are, hereafter, designated amorphous-CS, b-CS, and a-CS. The powders were sieved to 24–100 mesh (710–124 mm) before being used for the soaking experiments. The specific surface areas of the powders were calculated by the BET method based on N2 gas adsorption isotherms measured at 196 C (Autosorb-1, Quantachrome Co., FL, USA). 2.2. Soaking in simulated body fluid The simulated body fluid (SBF) was prepared by dissolving reagent-grade CaCl2, K2HPO4.3H2O, NaCl, KCl, MgCl2.6H2O, NaHCO3 and Na2SO4 in distilled water and adjusting the ion concentrations to be similar to those in human blood plasma (Table 1).6,7 The SBF solution was buffered at pH 7.25 with tris–(hydroxymethyl)-aminomethane [(CH2OH)3CNH2] and hydrochloric acid (HCl). The CaSiO3 powders were soaked in the SBF solution using the corresponding containers for 2 h, and 1, 5, 10, 15, 20, 25 and 30 days at a solid/liquid ratio of 1 mg/ml.10 The soaking experiment was carried out in a shaking bath maintained at 36.5oC (human body temperature). After the pre-selected soaking time, the powders were filtered off, rinsed in distilled water Table 1 Ion concentrations of simulated body fluid and human blood plasma
2.3. Analysis of hydroxyapatite (HAp) formation The soaked and dried CaSiO3 powders were crushed in an agate mortar to fine powders and investigated by X-ray diffractometry (XRD; Geigerflex, Rigaku Co., Tokyo, Japan) using monochromated CuKa radiation and Fourier transform infrared spectroscopy (FTIR; 8000PC, Shimadzu Co., Kyoto, Japan). The microstructures of the soaked powders were observed by scanning electron microscopy (SEM; s-2050, Hitachi Ltd., Tokyo, Japan). The concentrations of Ca, Si, and P in the SBF solution after soaking the CaSiO3 powders were determined by inductively coupled plasma atomic emission spectroscopy (ICP; SPS1500VR, Seiko Instruments Inc., Tokyo, Japan). For these analyses, 5 ml of the spent solution was diluted to 50 ml with distilled water.
3. Results and discussion 3.1. Characteristics of CaSiO3 powders Fig. 1 shows the XRD patterns (a) and SEM micrographs (b) of the three CaSiO3 powders. The XRD pattern of the amorphous-CS showed a broad peak (halo) at about 30o 2, while the patterns of the b-CS and a-CS confirmed the presence of crystalline b-CaSiO3 and aCaSiO3, respectively. Fig.1(b) shows the surface of both the amorphous-CS and the b-CS to be rough, but the aCS was smooth, consisting of elongated a-CaSiO3 grains with a very thick glassy phase at the grain boundaries. The chemical composition of this grain boundary glassy phase is considered to be silica-rich, since the product of its partial devitrification (cristobalite) is observed at 22o 2 in the XRD pattern. The specific surface area was 12.2 and 7.9 m2/g in the amorphous-CS and b-CS, respectively, but was below the measurable value (1 m2/ g) in the a-CS. The above results indicate that the increased firing temperature caused the powder particles to densify, especially in the a-CS in which it was assisted by formation of a liquid phase. This resulted in a different microstructure of the a-CS powder particles and a decrease in their specific surface area. 3.2. Characteristics of the formed HAp
Ion concentration (mM) Na+
K+ Mg2+ Ca2+ C1
Simulated body 142.0 5.0 fluid Blood plasma 142.0 5.0
HCO3
HPO24
1.5
2.5
148.8
4.2
1.0
1.5
2.5
103.0 27.0
1.0
3.2.1. XRD and FTIR Figs. 2–4 show the XRD patterns and FTIR spectra of the amorphous-CS, b-CS, and a-CS samples before and after soaking in the SBF solution. The XRD and FTIR data were measured after crushing the soaked
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
513
Fig. 1. XRD patterns of (a) three CaSiO3 powders and (b) SEM micrographs of these as-prepared powders. b: b-CaSiO3 (low temperature phase); a: a-CaSiO3 (high temperature phase); ~: cristobalite.
Fig.2. (a) XRD patterns and (b) FTIR spectra of the amorphous-CS before and after soaking in SBF solution; : calcite; *: HAp.
powders; these data, therefore, show the bulk constituent phases in the samples and not just the surface. Table 2 shows the assignments of the FTIR absorption bands.
In the amorphous-CS, a crystalline XRD peak observed at 29.3o 2 [Fig. 2(a)] after soaking for 2 h was identified as calcite (CaCO3). This is thought to be caused by the crystallization of an amorphous CaCO3
514
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 3. (a) XRD patterns and (b) FTIR spectra of the b-CS before and after soaking in SBF solution. b: b-CaSiO3; *: HAp.
Fig. 4. (a) XRD patterns and (b) FTIR spectra of the a-CS before and after soaking in SBF solution. a: a-CaSiO3; ~: cristobalite; *: HAp.
constituent of the original amorphous-CS. The FTIR C–O absorption band corresponding to this amorphous CaCO3 was detected at about 1485 cm 1 in the spec-
trum of the amorphous-CS prior to soaking. HAp formation was however detected in the FTIR spectrum of the 2 h soaked sample [Fig. 2(b)] as a composite
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
515
Table 2 Assignment for FTIR absorption bands (cm 1) of the CaSiO3 powders after soaking in the SBF solution Absorption band (cm 1)
Band assignment
Amorphous-CS
b-CS
a-CS
Calcium hydroxyapatite17
Carbonate hydroxyapatite19
3429 1415–1500 1080 c 1093
3440 1415–1500 1093
3450 1415–1512 1093
3572 – 1087, 1046
No report – 1037.6
953 875 800 606 567 468
952 870 800 605 567 468
953 877 800 609 565 466
962 – No report 601 571 474
958.2 873.7 No report 602.3 563.9 464
absorption band at about 1080 cm 1, composed of the P–O stretch of HAp and the Si–O vibration of the silicarich phase, superimposed on the silicate absorption band of the starting powder (991 cm 1). A crystalline peak of HAp at 31.7 2 corresponding to the 211 reflection was clearly observed in the XRD pattern after 1 day soaking. After prolonged soaking the HAp peaks became the main constituent of the XRD patterns as the HAp content increased and completely covered the surfaces of the powders, but strong peaks of calcite were also observed. These calcite peaks must arise from the CaCO3 residue inside the powder grains because these were crushed after the soaking experiments and before the XRD measurements. Anyway, co-existence of CaCO3 in the amorphous-CS may be disadvantage point for biomedical application because it is considered to show negative effect in vivo. The presence of calcite was not observed in the other two samples because the carbonate was completely removed from the b-CS and a-CS during firing. In addition, the FTIR spectra show the characteristic bands of HAp at about 1093 (shifted from 1080 cm 1), 606, and 567 cm 1 and also show the absorption bands at about 875 and 1415–1500 cm 1 of carbonate which is thought to be incorporated into the HAp. These vibrational modes of carbonate are typical of carbonate groups substituting for the phosphate (PO34 ) and/or hydroxide (OH ) groups in the HAp structure.18 In the b-CS, the 211 peak of HAp started to appear in the XRD pattern [Fig. 3(a)] after 1 day soaking although the b-CaSiO3 peak at 30 2 was still observed even after prolonged soaking. This result indicates that b-CaSiO3 still remains in the powder even after the surfaces are completely covered by the layer of HAp. The intensities of the HAp peaks in the b-CS are lower than in the amorphous-CS, suggesting that the crystallinity of HAp on the b-CS may be lower than that on the amorphous-CS. The FTIR spectra of Fig. 3(b) show a combination of the P–O absorption bands of the HAp, the Si–O absorption bands of the silica-rich phase, and the C–O absorption bands of the carbonate group which occur in almost the same positions.
O–H stretching C–O stretching of CO23 substitution for PU34 and/or OH P–O stretching of HAp (u3) Si–O stretching of silica-rich phase P–O stretching of HAp (u1) C–O stretching of CO23 substitution for PO34 and/or OH Si–O–Si stretching (tetragonal)2 O–P–O bending of HAp (u4) O–P–O bending of HAp (u4) O–P–O bending of HAp (u2) and/or Si–O–Si bending14
The a-CaSiO3 peak at about 32 2 in the a-CS was broadened due to its overlap with the 211 peak of HAp after 1 day soaking [Fig. 4(a)]. The main XRD peaks in this sample gradually changed from a-CaSiO3 to HAp until the a-CaSiO3 peaks were no longer detectable in the samples soaked for prolonged times. At the same time, the intensity of the cristobalite peak at 2=22o increased with increasing soaking time, indicating that although the aCaSiO3 grains dissolved completely in the SBF solution, the cristobalite did not dissolve even after prolonged soaking. The intensity increase of cristobalite is therefore due to a relative condensation effect. These results are quite different from the other two samples, probably because of differences in the microstructure of the starting powders. The FTIR spectra of Fig. 4(b) showed a change from the silicate absorption bands of the starting a-CaSiO3 to the P–O absorption bands of HAp, superimposed on the Si–O absorption bands of the silica-rich phase, and the C–O absorption bands of the carbonate group (Table 2). These results suggest a mechanism of HAp formation on the CaSiO3 powders involving its initiation from a rapidly-formed silica-rich interlayer, followed by the formation of an amorphous calcium phosphate layer resembling HAp by reaction of Ca2+, PO34 and OH ions from the SBF solution. After the prolonged soaking, the amorphous layer grows and crystallizes to HAp with partial incorporation of CO23 which may be related to the low crystallinity and small crystallite size of the HAp. 3.2.2. Microstructure of the formed HAp Figs. 5–7 show SEM micrographs of the surfaces of the amorphous-CS, b-CS, and a-CS after soaking in the SBF solution for various times. In the amorphous-CS, tiny ball-like agglomerated particles thought to be HAp were observed on the surface of the sample soaked for 2 h (Fig. 5). At longer soaking times, the number of agglomerated HAp particles increases with the formation of tiny new ball-like HAp particles rather than by growth in the size of the HAp particles already present. The surfaces of the long-soaked samples are covered with these tiny ball-like HAp particles. The texture of this layer consists
516
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 5. SEM micrographs of the surfaces of amorphous-CS after soaking in the SBF solution for various times.
of loosely packed agglomerates presenting a rough surface. The formation of this HAp layer occurred not only on the surface of the amorphous-CS, but also on the surface of the polypropylene container. This phenomenon did not occur in the soaking experiments with bCS or a-CS. Fig. 8 shows a SEM micrograph (a) and XRD pattern (b) of the HAp formed on the surface of the container after soaking the amorphous-CS for 5 days. This HAp occurs as porous spherical particles composed of flake-like crystals. This unique texture gives rise to preferred orientation in the XRD pattern manifested as strong intensities of the 211, 222, and 322 HAp reflections. The high crystallinity of the HAp formed in this sample is also thought to be related to this phenomenon. Figs. 6 and 7 show the HAp formed on the surfaces of both b-CS and a-CS powders after 1 day soaking to
occur as a dense layer consisting of tiny ball-like particles. On soaking for longer times, the HAp formed on both the b-CS and a-CS grew both as a result of increasing particle size and also with the formation of new HAp particles on the surfaces of other particles. Corresponding to the differences in growth behavior of these three samples, the HAp layers formed on the b-CS and a-CS surfaces appear denser than those on the amorphous-CS surfaces. The size of the product HAp particles decreased in the order a-CS>b-CS>amorphous-CS. Although a strong cristobalite XRD peak was observed in the soaked a-CS [Fig. 4(a)], no cristobalite particles were found in the SEM photographs of these samples. This result indicates that the cristobalite observed by XRD came from the original powder and was not formed during the soaking experiment.
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
517
Fig. 6. SEM micrographs of the surfaces of b-CS after soaking in the SBF solution for various times.
The above results show that the differences in formation behavior and microstructure of the product HAp layer arise from differences in the microstructure, surface area, and CaSiO3 phase of the starting powders. 3.2.3. Concentration changes of the SBF solution Fig. 9 shows the change in concentration of Ca, Si, and P in the SBF solution measured by ICP after soaking the three CaSiO3 powders for various times. During the early stages of soaking, the Ca and Si concentrations in the SBF solution increase steeply with a corresponding steep decrease in the P concentration. This reflects the dissolution of the powder surfaces with a simultaneous uptake of P from the solution into the powders. The amorphous-CS showed the highest release of Ca into the SBF solution but a lower release of Si than
from the other two samples. It is considered that the firing temperature of 500 C was insufficient for complete reaction of the CaO and SiO2 components of the amorphous-CS to form CaSiO3. The unreacted CaO may then carbonate to amorphous CaCO3 by reacting with CO23 in the SBF solution as suggested by the FTIR spectrum of the amorphous-CS [Fig. 2(b)]. The unreacted CaO and/or amorphous CaCO3 in the amorphous-CS are thought to dissolve faster in the SBF solution than either the SiO2, b-CaSiO3 or a-CaSiO3 phases. The higher release of Ca from the amorphousCS caused a more rapid increase in the pH of the SBF solution, resulting in the faster precipitation of HAp with the consumption of the Ca2+, PO34 , and OH as reflected by the steep decrease in the P concentration of the solution.
518
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 7. SEM micrographs of the surfaces of a-CS after soaking in the SBF solution for various times.
Fig. 8. (a) SEM micrograph and (b) XRD pattern of the HAp layer formed on the surface of soaking container (PE bottle) after soaking with the amorphous-CS for 5 days.
After prolonged soaking, the Ca concentration is similar in the solutions from amorphous-CS and b-CS, but is higher in the a-CS solution. This is due to the complete dissolution of a-CaSiO3, but not the b-CaSiO3
and CaCO3 of the other samples. The Si and P concentrations also showed differences between the samples, for reasons which are not presently understood. After prolonged soaking, the Ca, Si, and P concentrations
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
519
became almost constant, indicating an apparent equilibrium state between the dissolution of the powder surfaces and the formation of HAp.
4. Conclusions 1. CaSiO3 powders with different amorphous and crystalline phases showed differences in their microstructure and specific surface area. Rough, porous particles composed of aggregated fine particles occurred in amorphous-CS and b-CS and had higher specific surface areas than the smooth densely aggregated particles of a-CS. 2. These differences in microstructure and specific surface areas of the starting CaSiO3 powders influenced the formation behavior, microstructure, and particle size of the HAp formed on their surfaces after soaking in SBF solution. The HAp layer formed on the amorphous-CS was of a rough and loose texture, consisting of uniformly sized tiny ball-like particles, whereas a denser layer of larger particle size was formed on the b-CS and a-CS. 3. Amorphous-CS showed the fastest formation of HAp due to the rapid release of Ca. This product was however highly contaminated with CaCO3 which crystallized from the amorphous CaCO3 present in the starting powder.
Acknowledgements A part of this work was financially supported by the Japan Society for the Promotion of Science (Project: Preparation of Advanced Inorganic Materials by Soft, Solution Processing). We are also grateful to Dr. K.J.D. MacKenzie for critical reading and editing of the manuscript. References
Fig. 9. Changes of Ca, Si, and P concentrations of the SBF solution after soaking with the CaSiO3 powders for various times.
1. Hench, L. L., Splinter, R. J., Allen, W. C. and Greenlee, T. K., J. Biomed. Mater. Res. Symp., 1971, 2, 117. 2. Hench, L. L. J., Am. Ceram. Soc., 1991, 74, 1487. 3. Hench, L. L. and Andersson, O., An Introduction to Bioceramics. World Scientific, Singapore, 1993. 4. Jarcho, M., Kay, J. F., Gumaer, K. I., Doremus, R. H. and Drobeck, H. P., J. Bioeng., 1977, 1, 79. 5. Kokubo, T., Shigematsu, M., Nagashima, Y., Tashiro, M., Nakamura, T., Yamamuro, T. and Higashi, S., Bull. Inst. Chem. Res., 1982, 60, 260. 6. Kokubo, T., J. Non-Cryst. Solids, 1990, 120, 138. 7. Kokubo, T., An Introduction to Bioceramics. World Scientific, Singapore, 1993. 8. Ducheyne, P., Raemdonck, W. V., Heughebaert, J. C. and Heughebaert, M., Biomaterials, 1986, 7, 97. 9. Santos, J. D., Jha, L. J. and Monteiro, F. J., J. Mater. Sci.: Mater. Med., 1996, 7, 181. 10. Falaize, S., Radin, S. and Ducheyne, P., J. Am. Ceram. Soc., 1999, 82, 969.
520
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
11. Kokubo, T., J. Ceram. Soc. Japan, 1991, 99, 965. 12. Kokubo, T., Kushitani, C., Ohtsuki, C., Sakka, S. and Yamamuro, T., J. Mater. Sci.: Mater. Med., 1993, 4, 1. 13. Cho, S. B., Miyaji, F., Kokubo, T., Nakanishi, K., Soga, N. and Nakamura, T., J. Biomed. Mater. Res., 1996, 32, 375. 14. Kim, C. Y., Clark, A. E. and Hench, L. L., J. Non-Cryst. Solids, 1989, 113, 195. 15. Cho, S. B., Nakanishi, K., Kokubo, T., Soga, N., Ohtsuki, C. and Nakamura, T., J. Biomed. Mater. Res., 1996, 33, 145. 16. LeGeros, R. Z., Trautz, O. R., LeGeros, J. P. and Klein, E., Bull. Soc. Chim. France, 1968, 1712. 17. Fowler, B. O., Inorg. Chem., 1974, 13, 194.
18. Arends, J. and Davidson, C. L., Calcif. Tiss. Res., 1975, 18, 65. 19. Vallet-Regi, M., Romero, A. M., Ragel, C. V. and LeGeros, R. Z., J. Biomed. Mater. Res., 1999, 44, 416. 20. Siriphannon, P., Hayashi, S., Yasumori, A. and Okada, K., J. Mater. Res., 1999, 14, 529. 21. Siriphannon P., Kameshima, Y., Yasumori, A. and Okada, K. In Bioceramics 12, ed. H. Ohgushi, G. W. Hastings and T. Yoshikawa. World Scientific, Singapore, 1999, pp. 145–148. 22. Siriphannon, P., Kameshima, Y., Yasumori, A., Okada, K. and Hayashi, S., J. Biomed. Mater. Res., 2000, 52, 30. 23. Li, P., Ohtsuki, C., Kokubo, T., Nakanishi, K., Soga, N., Nakamura, T. and Yamamuro, T., J. Mater. Sci.: Mater. Med., 1993, 4, 127.
Formation of hydroxyapatite on CaSiO3 powders in simulated body fluid Punnama Siriphannona,1, Yoshikazu Kameshimaa, Atsuo Yasumoria, Kiyoshi Okadaa,*, Shigeo Hayashib a Department of Inorganic Materials, Tokyo Institute of Technology, O-okayama, Meguro, Tokyo 152-8552, Japan Research Institute of Materials and Resources, Mining College, Akita University, Tegatagakuen, Akita 010-8502, Japan
b
Received 10 January 2001; received in revised form 10 April 2001; accepted 21 April 2001
Abstract CaSiO3 powders were prepared from ethanol solutions of Ca(NO3)2.4H2O and Si(OC2H5)4 using NaOH as a precipitant. The resultant powders were heated at three different temperature regimes, (1) 500 C, (2) 500 and 1000oC and (3) 500 and 1400 C, to obtain the amorphous phase (amorphous-CS), low temperature phase (b-CS), and high temperature phase (a-CS) of CaSiO3, respectively. The different amorphous and crystalline phases exhibited different microtextures and specific surface areas of the powders. The rough, porous particles of amorphous-CS and b-CS have higher specific surface areas than the smooth, dense particles of a-CS. These CaSiO3 powders were soaked in a simulated body fluid (SBF) at 36.5 C for 2 h to 30 days. Formation of hydroxyapatite (HAp) was observed on the surfaces of all samples, but the formation behavior and microstructures were different, resulting the differences in microstructure and crystal structure of the starting powders as well as particle size and specific surface area. The HAp formed on the amorphous-CS was a loose porous layer consisting of uniformly-sized tiny ball-like agglomerated particles, while that formed on the b-CS and a-CS was a dense layer consisting of larger ball-like agglomerated particles. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Apatitie; Bioactive materials; CaSiO3 powders; Simulated body fluid (SBF)
1. Introduction During the last four decades, various materials known as ‘bioactive materials’ such as glasses,1 3 sintered hydroxyapatite,4 glass ceramics,5 7 composite materials,8 10 etc., have been synthesized and developed for medical applications. A significant characteristic of bioactive materials is their ability to bond with living bone through the formation of a hydroxyapatite (HAp) interface layer. The same HAp layer is reported to form on the surfaces of bioactive materials after soaking in a biomimetic system of simulated body fluid (SBF) which has ion concentrations similar to human blood plasma. The proposed mechanism of HAp formation involves the dissolution of calcium ions from the bioactive materials into the body fluid, with the simultaneous formation of * Corresponding author. E-mail address: [email protected] (K. Okada). 1 Present address: Department of Chemistry, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Chalongkrung Road, Ladkrabang, Bangkok 10520, Thailand.
a silica-rich interlayer on the material surfaces. The dissolution of the calcium ion increases the degree of supersaturation of the surrounding fluid with respect to HAp, while the silica-rich interlayer provides favorable sites for HAp nucleation by dissolving an appreciable amount of silicate ion.11 13 The nucleation and growth of the HAp layer proceeds by reaction of the calcium, phosphate, and hydroxide ions from the surrounding fluid,2,3,13 15 sometimes incorporating with carbonate or fluoride anions.2,14,16 19 a-CaSiO3 ceramics prepared from coprecipitated powders using NaOH as the precipitant have been proposed as new candidate bioactive materials by the present authors since they can induce the fast formation of an HAp layer after soaking in SBF solution.20 22 It is believed that the presence of calcium and silicate ions in the a-CaSiO3 ceramics plays an important role in the formation of HAp.23 The present study investigates the effect of the structure of the CaSiO3 powders on the formation behavior of HAp by soaking various amorphous and crystalline CaSiO3 powders in a SBF solution for varying periods.
0955-2219/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0955-2219(01)00301-6
512
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
and dried at room temperature. The soaking experiment was done only once for amorphous-CS, b-CS and a-CS, respectively.
2. Experimental 2.1. Preparation of CaSiO3 powders CaSiO3 powder was prepared by coprecipitation, using Ca(NO3)2.4H2O and Si(OC2H5)4 (TEOS) as the raw materials. These reagents were dissolved in 500 ml of ethanol, the concentration was adjusted to 0.2 mol/l, and the solution was stirred for 2 h. Precipitation was obtained by adding 0.33 mol/l NaOH solution to the solution. The precipitated gel was filtered, washed once with distilled water, dried in an oven at 100 C overnight, and calcined at 500 C for 2 h. The chemical composition of the as-prepared powder determined by X-ray fluorescence (XRF; RIX3000, Rigaku Co., Tokyo, Japan) was CaO 46.2, SiO2 53.4, and Na2O 0.4 mass% on a dryweight basis. The calcined powder (amorphous CaSiO3) was divided into three parts and subjected to different firing processes as follows; (1) fired at 1000oC for 2 h to obtain the low temperature crystalline phase b-CaSiO3 and (2) fired at 1400oC for 2 h to obtain the high temperature phase aCaSiO3. The resultant powders are, hereafter, designated amorphous-CS, b-CS, and a-CS. The powders were sieved to 24–100 mesh (710–124 mm) before being used for the soaking experiments. The specific surface areas of the powders were calculated by the BET method based on N2 gas adsorption isotherms measured at 196 C (Autosorb-1, Quantachrome Co., FL, USA). 2.2. Soaking in simulated body fluid The simulated body fluid (SBF) was prepared by dissolving reagent-grade CaCl2, K2HPO4.3H2O, NaCl, KCl, MgCl2.6H2O, NaHCO3 and Na2SO4 in distilled water and adjusting the ion concentrations to be similar to those in human blood plasma (Table 1).6,7 The SBF solution was buffered at pH 7.25 with tris–(hydroxymethyl)-aminomethane [(CH2OH)3CNH2] and hydrochloric acid (HCl). The CaSiO3 powders were soaked in the SBF solution using the corresponding containers for 2 h, and 1, 5, 10, 15, 20, 25 and 30 days at a solid/liquid ratio of 1 mg/ml.10 The soaking experiment was carried out in a shaking bath maintained at 36.5oC (human body temperature). After the pre-selected soaking time, the powders were filtered off, rinsed in distilled water Table 1 Ion concentrations of simulated body fluid and human blood plasma
2.3. Analysis of hydroxyapatite (HAp) formation The soaked and dried CaSiO3 powders were crushed in an agate mortar to fine powders and investigated by X-ray diffractometry (XRD; Geigerflex, Rigaku Co., Tokyo, Japan) using monochromated CuKa radiation and Fourier transform infrared spectroscopy (FTIR; 8000PC, Shimadzu Co., Kyoto, Japan). The microstructures of the soaked powders were observed by scanning electron microscopy (SEM; s-2050, Hitachi Ltd., Tokyo, Japan). The concentrations of Ca, Si, and P in the SBF solution after soaking the CaSiO3 powders were determined by inductively coupled plasma atomic emission spectroscopy (ICP; SPS1500VR, Seiko Instruments Inc., Tokyo, Japan). For these analyses, 5 ml of the spent solution was diluted to 50 ml with distilled water.
3. Results and discussion 3.1. Characteristics of CaSiO3 powders Fig. 1 shows the XRD patterns (a) and SEM micrographs (b) of the three CaSiO3 powders. The XRD pattern of the amorphous-CS showed a broad peak (halo) at about 30o 2, while the patterns of the b-CS and a-CS confirmed the presence of crystalline b-CaSiO3 and aCaSiO3, respectively. Fig.1(b) shows the surface of both the amorphous-CS and the b-CS to be rough, but the aCS was smooth, consisting of elongated a-CaSiO3 grains with a very thick glassy phase at the grain boundaries. The chemical composition of this grain boundary glassy phase is considered to be silica-rich, since the product of its partial devitrification (cristobalite) is observed at 22o 2 in the XRD pattern. The specific surface area was 12.2 and 7.9 m2/g in the amorphous-CS and b-CS, respectively, but was below the measurable value (1 m2/ g) in the a-CS. The above results indicate that the increased firing temperature caused the powder particles to densify, especially in the a-CS in which it was assisted by formation of a liquid phase. This resulted in a different microstructure of the a-CS powder particles and a decrease in their specific surface area. 3.2. Characteristics of the formed HAp
Ion concentration (mM) Na+
K+ Mg2+ Ca2+ C1
Simulated body 142.0 5.0 fluid Blood plasma 142.0 5.0
HCO3
HPO24
1.5
2.5
148.8
4.2
1.0
1.5
2.5
103.0 27.0
1.0
3.2.1. XRD and FTIR Figs. 2–4 show the XRD patterns and FTIR spectra of the amorphous-CS, b-CS, and a-CS samples before and after soaking in the SBF solution. The XRD and FTIR data were measured after crushing the soaked
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
513
Fig. 1. XRD patterns of (a) three CaSiO3 powders and (b) SEM micrographs of these as-prepared powders. b: b-CaSiO3 (low temperature phase); a: a-CaSiO3 (high temperature phase); ~: cristobalite.
Fig.2. (a) XRD patterns and (b) FTIR spectra of the amorphous-CS before and after soaking in SBF solution; : calcite; *: HAp.
powders; these data, therefore, show the bulk constituent phases in the samples and not just the surface. Table 2 shows the assignments of the FTIR absorption bands.
In the amorphous-CS, a crystalline XRD peak observed at 29.3o 2 [Fig. 2(a)] after soaking for 2 h was identified as calcite (CaCO3). This is thought to be caused by the crystallization of an amorphous CaCO3
514
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 3. (a) XRD patterns and (b) FTIR spectra of the b-CS before and after soaking in SBF solution. b: b-CaSiO3; *: HAp.
Fig. 4. (a) XRD patterns and (b) FTIR spectra of the a-CS before and after soaking in SBF solution. a: a-CaSiO3; ~: cristobalite; *: HAp.
constituent of the original amorphous-CS. The FTIR C–O absorption band corresponding to this amorphous CaCO3 was detected at about 1485 cm 1 in the spec-
trum of the amorphous-CS prior to soaking. HAp formation was however detected in the FTIR spectrum of the 2 h soaked sample [Fig. 2(b)] as a composite
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
515
Table 2 Assignment for FTIR absorption bands (cm 1) of the CaSiO3 powders after soaking in the SBF solution Absorption band (cm 1)
Band assignment
Amorphous-CS
b-CS
a-CS
Calcium hydroxyapatite17
Carbonate hydroxyapatite19
3429 1415–1500 1080 c 1093
3440 1415–1500 1093
3450 1415–1512 1093
3572 – 1087, 1046
No report – 1037.6
953 875 800 606 567 468
952 870 800 605 567 468
953 877 800 609 565 466
962 – No report 601 571 474
958.2 873.7 No report 602.3 563.9 464
absorption band at about 1080 cm 1, composed of the P–O stretch of HAp and the Si–O vibration of the silicarich phase, superimposed on the silicate absorption band of the starting powder (991 cm 1). A crystalline peak of HAp at 31.7 2 corresponding to the 211 reflection was clearly observed in the XRD pattern after 1 day soaking. After prolonged soaking the HAp peaks became the main constituent of the XRD patterns as the HAp content increased and completely covered the surfaces of the powders, but strong peaks of calcite were also observed. These calcite peaks must arise from the CaCO3 residue inside the powder grains because these were crushed after the soaking experiments and before the XRD measurements. Anyway, co-existence of CaCO3 in the amorphous-CS may be disadvantage point for biomedical application because it is considered to show negative effect in vivo. The presence of calcite was not observed in the other two samples because the carbonate was completely removed from the b-CS and a-CS during firing. In addition, the FTIR spectra show the characteristic bands of HAp at about 1093 (shifted from 1080 cm 1), 606, and 567 cm 1 and also show the absorption bands at about 875 and 1415–1500 cm 1 of carbonate which is thought to be incorporated into the HAp. These vibrational modes of carbonate are typical of carbonate groups substituting for the phosphate (PO34 ) and/or hydroxide (OH ) groups in the HAp structure.18 In the b-CS, the 211 peak of HAp started to appear in the XRD pattern [Fig. 3(a)] after 1 day soaking although the b-CaSiO3 peak at 30 2 was still observed even after prolonged soaking. This result indicates that b-CaSiO3 still remains in the powder even after the surfaces are completely covered by the layer of HAp. The intensities of the HAp peaks in the b-CS are lower than in the amorphous-CS, suggesting that the crystallinity of HAp on the b-CS may be lower than that on the amorphous-CS. The FTIR spectra of Fig. 3(b) show a combination of the P–O absorption bands of the HAp, the Si–O absorption bands of the silica-rich phase, and the C–O absorption bands of the carbonate group which occur in almost the same positions.
O–H stretching C–O stretching of CO23 substitution for PU34 and/or OH P–O stretching of HAp (u3) Si–O stretching of silica-rich phase P–O stretching of HAp (u1) C–O stretching of CO23 substitution for PO34 and/or OH Si–O–Si stretching (tetragonal)2 O–P–O bending of HAp (u4) O–P–O bending of HAp (u4) O–P–O bending of HAp (u2) and/or Si–O–Si bending14
The a-CaSiO3 peak at about 32 2 in the a-CS was broadened due to its overlap with the 211 peak of HAp after 1 day soaking [Fig. 4(a)]. The main XRD peaks in this sample gradually changed from a-CaSiO3 to HAp until the a-CaSiO3 peaks were no longer detectable in the samples soaked for prolonged times. At the same time, the intensity of the cristobalite peak at 2=22o increased with increasing soaking time, indicating that although the aCaSiO3 grains dissolved completely in the SBF solution, the cristobalite did not dissolve even after prolonged soaking. The intensity increase of cristobalite is therefore due to a relative condensation effect. These results are quite different from the other two samples, probably because of differences in the microstructure of the starting powders. The FTIR spectra of Fig. 4(b) showed a change from the silicate absorption bands of the starting a-CaSiO3 to the P–O absorption bands of HAp, superimposed on the Si–O absorption bands of the silica-rich phase, and the C–O absorption bands of the carbonate group (Table 2). These results suggest a mechanism of HAp formation on the CaSiO3 powders involving its initiation from a rapidly-formed silica-rich interlayer, followed by the formation of an amorphous calcium phosphate layer resembling HAp by reaction of Ca2+, PO34 and OH ions from the SBF solution. After the prolonged soaking, the amorphous layer grows and crystallizes to HAp with partial incorporation of CO23 which may be related to the low crystallinity and small crystallite size of the HAp. 3.2.2. Microstructure of the formed HAp Figs. 5–7 show SEM micrographs of the surfaces of the amorphous-CS, b-CS, and a-CS after soaking in the SBF solution for various times. In the amorphous-CS, tiny ball-like agglomerated particles thought to be HAp were observed on the surface of the sample soaked for 2 h (Fig. 5). At longer soaking times, the number of agglomerated HAp particles increases with the formation of tiny new ball-like HAp particles rather than by growth in the size of the HAp particles already present. The surfaces of the long-soaked samples are covered with these tiny ball-like HAp particles. The texture of this layer consists
516
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 5. SEM micrographs of the surfaces of amorphous-CS after soaking in the SBF solution for various times.
of loosely packed agglomerates presenting a rough surface. The formation of this HAp layer occurred not only on the surface of the amorphous-CS, but also on the surface of the polypropylene container. This phenomenon did not occur in the soaking experiments with bCS or a-CS. Fig. 8 shows a SEM micrograph (a) and XRD pattern (b) of the HAp formed on the surface of the container after soaking the amorphous-CS for 5 days. This HAp occurs as porous spherical particles composed of flake-like crystals. This unique texture gives rise to preferred orientation in the XRD pattern manifested as strong intensities of the 211, 222, and 322 HAp reflections. The high crystallinity of the HAp formed in this sample is also thought to be related to this phenomenon. Figs. 6 and 7 show the HAp formed on the surfaces of both b-CS and a-CS powders after 1 day soaking to
occur as a dense layer consisting of tiny ball-like particles. On soaking for longer times, the HAp formed on both the b-CS and a-CS grew both as a result of increasing particle size and also with the formation of new HAp particles on the surfaces of other particles. Corresponding to the differences in growth behavior of these three samples, the HAp layers formed on the b-CS and a-CS surfaces appear denser than those on the amorphous-CS surfaces. The size of the product HAp particles decreased in the order a-CS>b-CS>amorphous-CS. Although a strong cristobalite XRD peak was observed in the soaked a-CS [Fig. 4(a)], no cristobalite particles were found in the SEM photographs of these samples. This result indicates that the cristobalite observed by XRD came from the original powder and was not formed during the soaking experiment.
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
517
Fig. 6. SEM micrographs of the surfaces of b-CS after soaking in the SBF solution for various times.
The above results show that the differences in formation behavior and microstructure of the product HAp layer arise from differences in the microstructure, surface area, and CaSiO3 phase of the starting powders. 3.2.3. Concentration changes of the SBF solution Fig. 9 shows the change in concentration of Ca, Si, and P in the SBF solution measured by ICP after soaking the three CaSiO3 powders for various times. During the early stages of soaking, the Ca and Si concentrations in the SBF solution increase steeply with a corresponding steep decrease in the P concentration. This reflects the dissolution of the powder surfaces with a simultaneous uptake of P from the solution into the powders. The amorphous-CS showed the highest release of Ca into the SBF solution but a lower release of Si than
from the other two samples. It is considered that the firing temperature of 500 C was insufficient for complete reaction of the CaO and SiO2 components of the amorphous-CS to form CaSiO3. The unreacted CaO may then carbonate to amorphous CaCO3 by reacting with CO23 in the SBF solution as suggested by the FTIR spectrum of the amorphous-CS [Fig. 2(b)]. The unreacted CaO and/or amorphous CaCO3 in the amorphous-CS are thought to dissolve faster in the SBF solution than either the SiO2, b-CaSiO3 or a-CaSiO3 phases. The higher release of Ca from the amorphousCS caused a more rapid increase in the pH of the SBF solution, resulting in the faster precipitation of HAp with the consumption of the Ca2+, PO34 , and OH as reflected by the steep decrease in the P concentration of the solution.
518
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
Fig. 7. SEM micrographs of the surfaces of a-CS after soaking in the SBF solution for various times.
Fig. 8. (a) SEM micrograph and (b) XRD pattern of the HAp layer formed on the surface of soaking container (PE bottle) after soaking with the amorphous-CS for 5 days.
After prolonged soaking, the Ca concentration is similar in the solutions from amorphous-CS and b-CS, but is higher in the a-CS solution. This is due to the complete dissolution of a-CaSiO3, but not the b-CaSiO3
and CaCO3 of the other samples. The Si and P concentrations also showed differences between the samples, for reasons which are not presently understood. After prolonged soaking, the Ca, Si, and P concentrations
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
519
became almost constant, indicating an apparent equilibrium state between the dissolution of the powder surfaces and the formation of HAp.
4. Conclusions 1. CaSiO3 powders with different amorphous and crystalline phases showed differences in their microstructure and specific surface area. Rough, porous particles composed of aggregated fine particles occurred in amorphous-CS and b-CS and had higher specific surface areas than the smooth densely aggregated particles of a-CS. 2. These differences in microstructure and specific surface areas of the starting CaSiO3 powders influenced the formation behavior, microstructure, and particle size of the HAp formed on their surfaces after soaking in SBF solution. The HAp layer formed on the amorphous-CS was of a rough and loose texture, consisting of uniformly sized tiny ball-like particles, whereas a denser layer of larger particle size was formed on the b-CS and a-CS. 3. Amorphous-CS showed the fastest formation of HAp due to the rapid release of Ca. This product was however highly contaminated with CaCO3 which crystallized from the amorphous CaCO3 present in the starting powder.
Acknowledgements A part of this work was financially supported by the Japan Society for the Promotion of Science (Project: Preparation of Advanced Inorganic Materials by Soft, Solution Processing). We are also grateful to Dr. K.J.D. MacKenzie for critical reading and editing of the manuscript. References
Fig. 9. Changes of Ca, Si, and P concentrations of the SBF solution after soaking with the CaSiO3 powders for various times.
1. Hench, L. L., Splinter, R. J., Allen, W. C. and Greenlee, T. K., J. Biomed. Mater. Res. Symp., 1971, 2, 117. 2. Hench, L. L. J., Am. Ceram. Soc., 1991, 74, 1487. 3. Hench, L. L. and Andersson, O., An Introduction to Bioceramics. World Scientific, Singapore, 1993. 4. Jarcho, M., Kay, J. F., Gumaer, K. I., Doremus, R. H. and Drobeck, H. P., J. Bioeng., 1977, 1, 79. 5. Kokubo, T., Shigematsu, M., Nagashima, Y., Tashiro, M., Nakamura, T., Yamamuro, T. and Higashi, S., Bull. Inst. Chem. Res., 1982, 60, 260. 6. Kokubo, T., J. Non-Cryst. Solids, 1990, 120, 138. 7. Kokubo, T., An Introduction to Bioceramics. World Scientific, Singapore, 1993. 8. Ducheyne, P., Raemdonck, W. V., Heughebaert, J. C. and Heughebaert, M., Biomaterials, 1986, 7, 97. 9. Santos, J. D., Jha, L. J. and Monteiro, F. J., J. Mater. Sci.: Mater. Med., 1996, 7, 181. 10. Falaize, S., Radin, S. and Ducheyne, P., J. Am. Ceram. Soc., 1999, 82, 969.
520
P. Siriphannon et al. / Journal of the European Ceramic Society 22 (2002) 511–520
11. Kokubo, T., J. Ceram. Soc. Japan, 1991, 99, 965. 12. Kokubo, T., Kushitani, C., Ohtsuki, C., Sakka, S. and Yamamuro, T., J. Mater. Sci.: Mater. Med., 1993, 4, 1. 13. Cho, S. B., Miyaji, F., Kokubo, T., Nakanishi, K., Soga, N. and Nakamura, T., J. Biomed. Mater. Res., 1996, 32, 375. 14. Kim, C. Y., Clark, A. E. and Hench, L. L., J. Non-Cryst. Solids, 1989, 113, 195. 15. Cho, S. B., Nakanishi, K., Kokubo, T., Soga, N., Ohtsuki, C. and Nakamura, T., J. Biomed. Mater. Res., 1996, 33, 145. 16. LeGeros, R. Z., Trautz, O. R., LeGeros, J. P. and Klein, E., Bull. Soc. Chim. France, 1968, 1712. 17. Fowler, B. O., Inorg. Chem., 1974, 13, 194.
18. Arends, J. and Davidson, C. L., Calcif. Tiss. Res., 1975, 18, 65. 19. Vallet-Regi, M., Romero, A. M., Ragel, C. V. and LeGeros, R. Z., J. Biomed. Mater. Res., 1999, 44, 416. 20. Siriphannon, P., Hayashi, S., Yasumori, A. and Okada, K., J. Mater. Res., 1999, 14, 529. 21. Siriphannon P., Kameshima, Y., Yasumori, A. and Okada, K. In Bioceramics 12, ed. H. Ohgushi, G. W. Hastings and T. Yoshikawa. World Scientific, Singapore, 1999, pp. 145–148. 22. Siriphannon, P., Kameshima, Y., Yasumori, A., Okada, K. and Hayashi, S., J. Biomed. Mater. Res., 2000, 52, 30. 23. Li, P., Ohtsuki, C., Kokubo, T., Nakanishi, K., Soga, N., Nakamura, T. and Yamamuro, T., J. Mater. Sci.: Mater. Med., 1993, 4, 127.